|Publication number||US7483049 B2|
|Application number||US 10/006,444|
|Publication date||Jan 27, 2009|
|Filing date||Nov 20, 2001|
|Priority date||Nov 20, 1998|
|Also published as||US20030095186, WO2003044734A2, WO2003044734A3|
|Publication number||006444, 10006444, US 7483049 B2, US 7483049B2, US-B2-7483049, US7483049 B2, US7483049B2|
|Inventors||James A. Aman, William R. Haller|
|Original Assignee||Aman James A, Haller William R|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (61), Non-Patent Citations (23), Referenced by (65), Classifications (25), Legal Events (5)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The present invention is a continuation in part of U.S. patent application Ser. No. 09/510,922 filed on Feb. 22, 2000, now issued U.S. Pat. No. 6,707,487 B1, as well as U.S. patent application Ser. No. 09/197,219 filed on Nov. 20, 1998, now issued U.S. Pat. No. 6,567,116.
The present invention relates to machine vision systems for tracking the movement of multiple objects within a predefined area or volume.
Several systems currently exist in the commercial marketplace for tracking the movement of one or more objects within a limited area. Such systems are routinely used to follow the movement of human subjects for a range of purposes including medical analysis, input for lifelike animation, as well as sports measurement. The following companies provide machine vision-based motion analysis systems:
Each of these systems shares many drawbacks, especially when considered for use in a live sporting environment such as a sporting contest. Some of these drawbacks are as follows:
The present inventors have addressed many of these drawbacks in their co-pending applications entitled:
Each of these patent applications is hereby incorporated by reference into the present application.
In these patents applications, the present inventors describe various aspects of a multiple object tracking system that functions in general to track many types of objects but that is especially constructed to track athletes during a live sporting event such as an ice hockey game. These patent applications teach at least the following novel components:
Also currently existing in the commercial marketplace are the following important components:
The present inventors have described in their co-pending applications many useful component and system solutions to the problems that are inherent within the existing systems. Additional components as described above also exist. It is possible to create several different and yet effective machine vision systems for tracking multiple moving objects based upon the novel components disclosed within the present application and four co-pending applications. What is needed is an understanding of how all of these teachings can be combined to form several different machine vision systems, each with their own novel optimizations.
In addition to the aforementioned machine vision system solutions to multi-object tracking it should be noted that at least two other companies are attempting to provide systems for similar purposes. Both Trackus, a Massachusetts-based company, and Orad, an Israeli-based company, are attempting to develop real-time “beacon”—based tracking technology for sporting events. Orad has produced a working system to follow horse racing; Trakus is the only company currently attempting to follow players in an ice hockey game. While Orad's solution is essentially similar at the highest levels, the technology will be explained based upon information gathered concerning Trakus.
Trakus' solution includes a microwave based transmitter and receiver that will track a single point within the helmet of each player. There are many deficiencies with this proposed solution as compared to machine vision systems in general and the novel teachings of the present inventors in particular. One of the most important distinctions is the “active” and potential harmful nature of the microwave technology. If used for tracking youth sports, it is anticipated that the average parent would balk at the idea of placing even a low-power microwave device into the helmet of their child. Furthermore, there are significant reflection problems due to the hard interior surfaces of a hockey arena that must be resolved before this technology can effectively track even a single point (the helmet) on every player on both teams. As already discussed, machine vision-based systems employ “passive” markers that are capable of tracking 14 or more points (the head and every major joint) on every player in real time. The present invention furthermore uniquely teaches a system that can also track game equipment and the puck, devices that have surfaces that cannot be substantially altered by the normal size of traditional markers.
For the purposes of disclosing the novel teachings of the present invention, the exemplary application of following the motion of hockey players, their equipment, and the game puck in a live sporting event will be used to represent multi-object tracking.
In order to create an optimal multi-object tracking system, the present inventors have focused on four major factors as follows:
A. The desired tracking information to be determined by the system;
B. The characteristics of the objects to be tracked;
C. The characteristics of the tracking environment; as well as
D. Traditional engineering goals.
A. With respect to the desired tracking information to be determined by the system, the following characteristics were considered:
B. With respect to the objects to be tracked, the following characteristics were considered:
C. With respect to the tracking environment, the following characteristics were considered:
D. With respect to traditional engineering goals, the following characteristics were considered:
An optimized system design such as disclosed in the present application must consider many of the above-stated questions and criteria simultaneously. However, for the sake of consistency, each of the pertinent questions and criteria listed above will be considered in order.
Therefore, referring first to the characteristics of the desired tracking information to be determined by the system, the following is taught.
A.1. Is the desired representation to be visual for display only or a mathematical model for measurement and rendering?
It is preferable that the tracking system creates a mathematical model of the tracked players and equipment as opposed to a visual representation that is essentially similar to a traditional filming and broadcast system. A mathematical model allows for the measurement of the athletic competition while also providing the basis for a graphical rendering of the captured movement for visual inspection from any desired viewpoint. Certain systems exist in the marketplace that attempt to film sporting contests for multiple viewpoints after which a computer system may be used to rotate through the various overlapping views giving a limited ability to see the contest from any perspective. All of the aforementioned machine vision companies, such as Motion Analysis and Vicon, generate a mathematical model.
This requirement of creating a mathematical model of human movement necessarily dictates that at least one precise location on the human body be identified and followed. For beacon-based systems such as Trakus or Orad, following the beacon's signal provides both identity and object location. To be implemented in a machine vision system, this further implies that each of these locations remain substantially in view of two or more tracking cameras at all times. Furthermore, the use of markers strategically placed upon the players can greatly simplify video frame analysis, as the markers become consistent center points that reduce the need for detailed edge detection techniques as well as identification and weighted averaging of player surfaces. This marker technique is implemented by all of the vision-based systems such as Motion Analysis and Vicon and is preferred by the present inventors. As will be discussed, the exact choice of the type, shape, size, and placement of the preferred markers is significant to some of the novel functions of the present invention and is different from existing techniques.
A.2. Is two or three-dimensional information preferred?
Three-dimensional information provides the ability to generate more realistic graphical renditions and to create more detailed statistics and analyses concerning game play. All of the aforementioned machine vision companies such as Motion Analysis and Vicon attempt to generate three-dimensional data while the beacon-based Trakus and Orad only generate two-dimensional information.
This requirement of creating a three-dimensional mathematical model of game play further dictates that at least the major joints on a player are identified and tracked. For instance, the player's helmet, shoulders, elbows, writs, torso, waist, knees, and feet are all beneficial tracking points for a 3D model. This informational goal in practice precludes beacon-based systems such as Trakus and Orad since it would require a significant number of beacons to be placed onto each player, often times in locations that are not convenient. Furthermore, each beacon will create additional signal processing and, given current state of the art in microwave tracking, the system could not function quickly enough to resolve all of the incoming signals.
For vision-based systems such as Motion Analysis and Vicon, significant difficulties also begin to present themselves in consideration of the requirement that each joint be in view of at least two cameras at all times. As players move about and change their body positions, individual joints can easily be lost from view (inclusion) or take positions that from a given camera's “flat 2D” perspective appear to make them a part of a different player. This is especially true in light of the small spherical markers used by existing systems that are not in view from every possible rotational angle of the joint. What is needed is a vision-based tracking system that can identify and track players with one set of cameras and then automatically direct a second set of cameras to adjust their views so as to minimize these inclusions. What is further needed is a tracking system that employs markers that can cover a much larger area, for instance all the way around an elbow rather than a single point or set of three points upon the elbow, while at the same time remaining less obtrusive.
A.3. Is object orientation required in addition to location, velocity, and acceleration?
Especially in the sport of ice hockey, it is important to know the orientation of the player as a whole since, for instance, they can be skating in a certain direction either facing forwards or backwards. Furthermore, in any motion system designed to follow the movement of a human joint that has multiple degrees of freedom, it is necessary to determine the orientation of the joint, not merely its position in order to create an accurate mathematical model. This requirement exceeds the capacity of beacon-based systems such as Trakus and Orad since the emitted signals are uniformly omni-directional and therefore cannot be used to determine rotation about the transmitting axis. For vision-based systems such as Motion Analysis, Vicon and the present invention, this requirement significantly favors the use of markers over edge detection of player surfaces. This is because the markers have clear center points and could be placed in a triangular format that is a traditional method for orientation detection. The present inventors prefer larger markings that form shapes such as an oval, circle, or triangle over the placement of spherical markers as is currently practiced.
A.4. Must this information be collected and available in real time?
The ability to capture and analyze images, convert them into a 3-D mathematical model, and then dynamically render a graphic representation along with quantified statistics in real time offers significant opportunities and challenges. For beacon-based systems such as Trakus and Orad, there is a single set of receiving towers throughout the arena that must process as many beacons as required to follow all of the players to be tracked. Essentially, the entire set of towers is necessary to cover the rink for even a single player.
Additionally, each player's beacon signal will be picked up by each tower and then must be compared across all towers to perform the location function. As more players are added, they will create additional processing for each tower. Hence, the total number of players is limited by the capacity of a single tower to process its received signals in real time. While such beacon-based systems may be scalable by playing area, therefore simply adding additional towers will cover additional area, they are not scalable by the number of players. Hence, for more players within a given area, one cannot simply add more towers. These systems are inherently “player bound” as opposed to “area bound.”
For vision-based systems such as Motion Analysis and Vicon, the goal of real-time information significantly challenges system capacity especially when combined with the need to track multiple players. This requirement limits the amount of image processing time available to handle ambiguities created in the data set by the inclusions that occur when players overlap, or bunch up, within a given camera view. Similar to the beacon-based systems, these camera systems can be said to be scalable by playing area since each additional playing volume requires the uniform addition of fixed perspective cameras. However, they too are “player bound” since any number of players may at a single time end up in any given playing volume creating a large, “included” data set that cannot be sufficiently processed in real time if at all without human assistance to clarify ambiguities.
The present inventors prefer to separate the function of “area” tracking for 2D movements and identification from “player” tracking for the full 3D data set. Hence, a matrix of overhead (X, Y) cameras follows the movement of any number of players per single camera across a single playing area. Since the number of markers tracked from the overhead view is limited to primarily the helmet and shoulders, the total number of marks, even for a large number of players is still trackable. Furthermore, due to the overhead view and the tendency for players to remain upright, it is expected that there will be minimal instances of helmet or shoulder “overhead inclusions.” In addition to being tracked in 2D space by the overhead cameras, each player will also be followed by at least two and preferably four dedicated pan, tilt, and zoom perspective cameras. These dedicated, movable, perspective cameras will be automatically directed by the known (X, Y) location of each player as determined by the overhead cameras. This combination of novel techniques provides for scalability by both tracking area and player. Hence, to cover more area simply add one or more overhead “area” tracking cameras while to cover more players simply add additional sets of one or more dedicated movable “player” tracking cameras.
A.5. What is the acceptable accuracy and precision with respect to this information?
The speed of a hockey puck can approach 100 mph while a player may be skating upwards of 25 mph. As was previously described in detail in the present inventors' co-pending application for a Multiple Object Tracking System, minimum capture rates of 40 frames per second for low-end commercial video cameras will follow pucks at a maximum of 3.7 feet per frame and players at 0.9 feet per frame. While this is sufficient to create a realistic mathematical model of game movement, faster cameras with between 2 to 4 times the capture rates are readily available. The incorporation of faster cameras also requires the use of faster computers and software algorithms to locate and identify markers and therefore players and equipment within the allotted fraction of a second. Currently, the beacon-based Trakus captures data at 30 locations per second.
A.6. Will the system be required to identify the objects as well as tracking them?
The ability to identify the tracked objects is critical and poses a more difficult problem for the video-based systems versus the beacon systems. With a beacon system, the transmitted signal can easily contain a uniquely encoded value that serves to differentiate each player. For a video-based system, which uses a single tracking energy and passive markers, the encoding must be accomplished via some unique arrangement of markings. For vision based systems such as Motion Analysis and Vicon, the current practice is to consider the unique “constellation” of markers as placed upon each players body to form an encoding for that player. This thinking is predicated on the idea that no two individuals have exactly the same body shape and hence the markers will always be in at least slightly different configurations. If two players did end up with “constellations” that were too similar in configuration to be accurately differentiated, then one or more additional markers would be added to at least one of the players to sufficiently differentiate the one from the other. It should be noted, that with this strategy, the majority of the markers play a dual role of both joint and body segment tracking on an individual basis and player identity on a collective basis.
A further practical limitation of this technique is its real-life implementation with non-sophisticated system operators. In other words, at the most difficult youth sport level, the ideal tracking system must be operable by minimally trained lay people such as coaches and parents. Placing these markers upon the player's body in such a way as to guarantee a unique constellation for all 16 to 20 players on a team will be overly restrictive. What is needed is a simple approach to placing a minimum of a single marker upon the player that is assured to provide a unique identification during game tracking.
The present inventors prefer to isolate player identification to those markings that can be viewed strictly from the overhead cameras, i.e., the “top surface” of the player's body. This minimizes the occasions when the markers that are relied upon for identification are hidden (included) from camera view based upon player bunching. In the preferred embodiment, the markings placed on the player's helmet and possibly also on their shoulders form a uniquely identifiable pattern, similar in form to bar coding. The decision as to whether the shoulder markers need to be included in the unique pattern depends mainly upon the total area provided by the helmet alone and the number of pixels per inch resolution of the viewing cameras. For effective machine vision, any mark to be “seen” must be picked up by preferably two pixels. Depending upon the pixel per inch resolution of the camera configuration, these two pixels will represent a different sized area. In practice, due to the expected movement and rotation of the helmet that is inherently not a flat surface, it is preferred to have many more pixels per inch of marking for accurate recognition. Additionally, it is anticipated that there may be between thirty to forty players and game officials on the ice over the course of a given game. Each one of these players and officials must have a unique tracking pattern on their helmet. Hence, if there is not enough room on the helmet to include an easily recognizable, unique encoding for each player, the shoulders can be used to augment the coding scheme.
This novel approach provides the opportunity to optionally separate the player identification process from the tracking process. Since the players are all expected to enter and exit the playing area via a limited passageway and furthermore to collect just outside the playing area on team benches, it is anticipated that these will present ideal areas to first ascertain and then reconfirm unique player identities. Once the players are identified as they first enter the playing area through the passageway, they can then simply be tracked as they move throughout the playing area. Should an ambiguity arise between two players during game play, their identity can be sorted out once they return to the team benches.
This separation of player identification from movement tracking opens up the possibility of different camera and lens configurations in the “identification areas” versus the “tracking areas.” Hence, the identification cameras could for instance be focused on a much narrower field of view since the passageway and team bench are considerably smaller in area than the playing surface. By narrowing the field of view, the resultant pixels per square inch of “id pattern” will be increased thereby reducing the necessary size of this unique marking. Other possibilities are anticipated such as the creation of a unique marking that is detected in a first select frequency such as IR and an overlapping tracking mark that is detected in a second select frequency such as UR. (These frequencies could easily be reversed with no change to the novel functionality.) In this embodiment, the “identification areas” will be viewed with IR cameras while the “tracking areas” will be viewed with UV cameras. It should be noted that both the IR id mark and the UV tracking mark would be non-visible to the players and viewing audience. Furthermore, these separate “identification areas” will also facilitate tracking the limited movement of the players until they enter the playing surface, i.e. “tracking area.”
A.7. Once an object has been identified, can this identity be lost during tracking?
The answer to this question is entirely driven by the selected tracking and identification technique. For instance, with beacon-based identification such as Trakus and Orad, the identity is only lost if the unique player signal is lost. For traditional vision-based tracking such as Motion Analysis and Vicon, player identity can be easily lost based upon a number of factors. These factors include the number of cameras viewing a given playing volume, the pixel per inch resolution for the allowed marker size as well as the number of expected inclusions due to player bunching. The final factor is randomly variable throughout a given tracking session. The preferred embodiment of the present invention has separated the identification of each player into a 2D overhead viewing function that strictly focuses on the “top surface” of each player. Since the number of inclusions of player's helmets is expected to be minimal, especially from the overhead view is expected to be minimal, the preferred embodiment will not experience the same difficulties in maintaining player identity as traditional systems.
B. Referring next to the characteristics of the objects to be tracked, the preferred and alternate novel embodiments must consider at least the following factors:
B.1. Is the object rigid or flexible?
The majority of objects (i.e., the players) are flexible while some objects are rigid such as the sticks and puck. Flexible objects such as players create difficulties for vision-based tracking systems because their form is constantly changing. These changes create greater opportunities for inclusions (blocked markers) and for misinterpretation when multiple players wearing multiple markers are each partially in view from any given camera. For beacon-based systems such as Trakus and Orad, the only location being tracked on the player is the helmet that is itself rigid. Due to the nature of the beacon itself, it is immaterial whether the object is flexible since the beacon's signal will transmit through the player's body and most equipment. For machine vision systems such as Motion Analysis and Vicon, this restriction significantly adds to the difficulty of keeping any given marker in the view of two or more cameras at a single instant, especially in light of the multiple player requirements.
Due in part to the infinite variations of marker images that can be created by any number of players holding their bodies in any number of positions creating any number of inclusions, all within a single camera field of view, the present inventors favor the novel teachings of the present invention. More specifically, by isolating player identification to markings placed upon the player's helmet that is rigid and substantially in view of the overhead cameras at all times, the flexibility of the player's body is removed as a negative factor for the more complex purpose of identification. Additionally, the use of larger ink markings provides for various marker shapes such as circles around the elbows, torso, knees or ankles or strips that can run down a player's arms, legs, torso or stick, etc. This in turn allows the present invention to greatly reduce the chances for total inclusion as the player's flex individually and bunch together. And finally, the separation of overhead (X, Y) “area” tracking that automatically directs the dedicated pan, tilt and zoom “player” tracking cameras, which in turn collect the 3D information, further minimizes the inclusions created by multiple flexible objects.
B.2. In how many degrees of freedom will the object be moving?
Similar to the difficulties created by the flexibility of a player, the full six degrees of freedom within which a player and or their limbs may travel creates increased challenges for vision-based solutions such as Motion Analysis and Vicon. It should first be noted that beacon-based systems such as Trakus and Orad are not impacted by the freedom of the player to move their helmet in any of the six possible directions from the current location. Of course, the beacon solution itself has a much greater shortcoming since it is in practice limited to a single beacon per player.
Returning to the vision systems such as Motion Analysis and Vicon, the current practice is to place small spheres ranging in size from approximately ⅛″ to 1″ in diameter upon the various joints and locations to be tracked on any given player. Often, these small spheres are themselves held out away from the players body on a short stem of approximately ½″. This configuration is uniformly adopted by the current 3D vision-based tracking systems and represents a significant system limitation. The purpose of the spherical shape of the marker is to ensure that a maximum image size is created irrespective of the location of the marker (and its associated body part) with respect to the viewing camera. Hence, no matter what the angle of view, the marker will always show up as a circle unless the view is in some way blocked. The purpose of the stem is to hold the sphere out away from the body thereby reducing the circumstances of partial inclusion created by the body surface to which the marker has been attached. While this technique has worked well for “controlled” situations, these markers are unacceptable for “live” events.
What is needed is a method of marking the player's joints that will be substantially visible from two or more cameras from any view point despite the six degree freedom of player movement. Given this requirement, the present inventors prefer the novel technique of using a much larger “surface area” marking. This essentially maximizes the marker view by both increasing the size of the mark as well as continuously surrounding a joint or body part such as the elbows, torso, knees or ankles. By both enlarging the mark and surrounding the body part, the ability for that part to been seen from any angle while it moves in six degrees of freedom is significantly increased.
B.3. How fast will the objects be moving?
As was previously stated, the speed of a hockey puck can approach 100 mph while a player may be skating upwards of 25 mph. When players are spinning or turning at or near their full speed of travel, the combined joint movement speed can present problems to tracking systems with lower sampling rates. For beacon-based systems such as Trakus and Orad, which take approximately 30 locations per second, these speeds would present a significant challenge if the beacons could be placed upon a puck or even a player's wrist. For vision-based systems such as Motion Analysis, Vicon and the present invention, camera and video capture technology can presently handle upwards of 240 frames per second with acceptable 1 megapixel resolution.
B.4. Are there multiple objects to be tracked and, if so, how will this impact the tracking method?
As has been previously suggested, it is necessary to track the motion of at least 10 players and 3 game officials in the playing area for the sport of ice hockey. Including the players' benches, the total number of players and officials can exceed 40. For other sports such as American football, the total players and officials can exceed 60, over 20 at a time on the playing surface. This requirement for ice hockey alone has pushed the limits of beacon-based systems such as Trakus and Orad that are simply attempting to track a single point, the helmet, for each player. For vision-based systems such as Motion Analysis and Vicon, this requirement also presents a significant challenge. All of the existing vision tracking systems first place fixed field of view cameras just outside the tracking area. These cameras are strategically placed in an attempt to keep the maximum number of markers within view of at least two cameras at all times even as these markers travel about in six degrees of freedom. The prospect of viewing multiple players within a given volume further challenges the strategic placement of these fixed cameras. As more and more players randomly congregate within any given volume, more and more of their respective markers will be blocked, from any given camera's field of view. Since the sum total of a player's markers is used in and of itself for player identification, as markers are lost from view this not only jeopardizes the accurate tracking of an individual body part, but also increases the incidence of improperly identified players.
The present inventors prefer the novel approach of controllably directing one or more movable pan, tilt and zoom cameras to follow each individual player based upon the first determined (X, Y) location as obtained via the overhead cameras. In this way the system's “total field-of-view” is dynamically recreated and maximized for player and marker visibility. The preferred embodiment includes movable cameras that are not necessarily unconditionally dedicated to follow a single player. It is anticipated that, given the known and projected locations of each player in the tracking area, it may be beneficial to dynamically switch one or more cameras away from one or more players onto one or more other players.
B.5. Are there any restrictions or safety considerations regarding the type of electromagnetic energy used to track the objects?
For safety requirements the properties of both wavelength and flux must be considered apart and in combination. The beacon-based tracking implemented by Trakus employs microwaves similar to cell phone technology. These frequencies of energy can be harmful in larger amounts and, for this reason, Trakus pulses their transmitter signals so as to reduce the average energy exposed to the player through the helmet. However, the present inventors see at least the potential for perceived harm since the players, which in the case of recreational hockey are youths, will be exposed to higher dosages of this energy as they tend to bunch together during play or at least in the team bench area.
In the case of vision-based systems such as Motion Analysis or Vicon, the primary tracking energy is visible, such as red light. While these energies are not harmful to the players, they do restrict the use of these systems to controlled settings where it is acceptable for the athletes and viewers to see the markers and their reflections. Motion Analysis, Vicon and other providers also offer the choice of working in near IR frequencies. These IR frequencies are safe for human exposure. The ring lights used to illuminate the tracking area in these systems only emit IR energy, however, and as such are not noticeably active on to the casual viewer. There is some concern that an uninformed individual could stare up at the camera with its IR ring light and receive an overexposure of IR energy possibly damaging the retina of the eye. For this reasons, these ring lights come with additional small LED's that emit a visible light cue when the IR light is on, essentially alerting people not to stare. This may not be a sufficient mechanism for the uniformed casual viewer. When tracking in IR energy, the present inventors prefer a light source that emits both visible and IR light since the visible light will act as a natural cue to dissuade the observer from staring.
Also taught in the preferred embodiment is the novel use of UV energy for tracking purposes. UV energy is broken into three types; UVC, UVB and UVA. UVC and UVB are generally thought to be the most damaging while UVA is considered to be biologically safe and much more prevalent in the atmosphere than UVB. UVA is typically considered to be those wavelengths between 315 to 400 nm. Visible light begins as 400 nm. The preferred embodiment employs tracking frequencies centering around 390 nm, just shy of visible light. This is an ideal wavelength since there are low-cost industrial digital cameras whose responsivity curves peak at or near this frequency. Furthermore, the metal halide lamps that are often found in hockey rinks generate a significant amount of UVA energy that could be used as an energy source.
B.6. What are the physical space limitations on the objects for any marker or beacon-based tracking system?
To employ beacon-based tracking such as Trakus and Orad for contact sports it is preferable to embed the beacon somewhere in the player's equipment. Due to several other drawbacks, these beacon systems have been limited to tracking a single point that was centrally chosen to be the helmet. If other considerations were permitting and additional beacons could be tracked per player, then the requirement to embed beacons into the equipment in order to avoid injury possible from player contact would itself becoming a significant problem.
For vision-based systems such as Motion Analysis and Vicon, the markers must reside outside of the player's exposed surface in order to receive and reflect the tracking energy. Furthermore, as was previously mentioned, these markers have been constructed as spheres so as to maximize the reflected image independent of marker angle to camera. These spheres strategically placed onto various joints and body parts are impractical for live, uncontrolled settings such as a full-contact sporting event. The preferred embodiment of the present invention specifies the use of reflective ink or paint that can be applied to the various substrates such as plastic, wood, fabric, leather, rubber, etc. and will add minimal thickness. Unlike to current practice of using visible spheres of limited diameter that must protrude from the player to maximize reflection, the preferred technique uses non-visible ink to place large markings on the various joints and body parts. Depending upon the transmisivity to the fabric being worn by the players, the ink itself could be placed on the inside of this material and receive and reflect the non-visible tracking energy directly through the jersey. The end result of this type of preferred marking is to overcome any physical space limitation considerations.
B.7. What is the natural reflectivity of the various background surfaces to the different potential tracking energies?
The reflectivity of the background surfaces, and for that matter the “non-tracked” foreground surfaces, represents the most significant source of system noise. For beacon-based systems such as Trakus and Orad, the microwaves emitted by the players' transmitters will reflect off hard surfaces such as metal and concrete. Due especially to the enclosed environment of a hockey arena, these reflections will continue to bounce between the background surfaces within the tracking volume as they slowly attenuate, thus contributing to significant noise problems.
For vision-based systems such as Motion Analysis and Vicon, reflectivity issues are limited to the individual camera's field of view. Hence, since the cameras are primarily focused on the playing surface and surrounding boundaries, the reflection issues are already diminished over beacon-based systems. However, for the sport of ice hockey, the playing surface is ice that tends to be highly reflective of visible, IR and UV energies. In the case of visible red light as preferred by Motion Analysis and Vicon, these reflections are typically handled via software processing. During system calibration, the images captured by each camera are reviewed for possible ice surface reflections especially from the ring lights affixed to the tracking cameras. Any such undesirable signals are “mapped” out via a software tool that essentially informs the image analysis to ignore any and all data capture from those coordinates. Of course, should valid marker data coincidentally show up at that same location during a live filming, it would be ignored as well.
In the preferred embodiment, absorbent compounds are applied to the various background and “non-tracked” foreground surfaces. For instance, if the tracking energy is UV, traditional UV absorptive compounds as are well known in the commercial marketplace can be used to absorb stray UV energy. The entire player including their jersey, helmet, pads, gloves, stick, etc. as well as the ice surface, boards and glass can all be first treated with one or more UV absorptive compounds. This novel technique significantly reduces system noise and essentially makes everything but the subsequently applied markers invisible to the tracking cameras. Similar techniques are possible if IR is used as the tracking energy.
C. Referring next to the characteristics of the tracking environment, the preferred and alternate novel embodiments must consider the following questions:
C.1. Is the setting “live” or “controlled”?
Tracking players in a live sporting event such as ice hockey adds a major restriction to the system requirements. Namely, the system must be as unobtrusive as possible for both the players and the viewing audience. In the case of the beacon systems from Trakus and Orad, the transmitters are hidden and the tracking signal is non-visible. The only possible drawback is that receiving towers must be placed throughout the arena. Careful attention must be made to keep these towers from blocking the view of nearby fans.
With respect to vision-based systems such as Motion Analysis and Vicon, there are two major types of restrictions present in their approach. First, the tracking energy itself is typically within the visible spectrum (e.g. red light). This energy is necessarily also visible to the players and audience. Second, as previously discussed, the markers have substantial size. Their size makes them noticeable to players and fans and potentially dangerous to the players since they could apply additional pressure to their bodies upon contact with other players, the ice surface or the boards. Furthermore, these markers could easily be dislodged creating at a minimum the stoppage of game play and or worse a potential injury hazard. Motion Analysis and Vicon, as well as other vision tracking providers, offer an option to work in IR light. This change eliminates the direct interference of visible light and its reflections off the ice surface with player and fan vision.
However, the markers used by these types of systems have an additional problem other than their size. Specifically, they are constructed of a spherical material preferably a foam ball, that has been covered with a retroreflective tape. The tape itself is greyish in visible color. These tapes are typically made of a material referred to as “cube cornered retroreflectors” but could also be produced with microspheric devices. In either case, current technology provides for optically transparent bodies that are coated on their undersides with a reflective material such as aluminum or silver, essentially forming a mirror. The additional problem is the broad band frequency response of these microscopic mirrors that includes visible light as well as UV and IR. Hence, even if the tracking energy is switched to IR, other ambient visible frequencies will themselves retroreflect off these markers back towards the players and fans who are near the light source. As players move about, these spheres will have a tendency to fluctuate in brightness as they pass through various lighting channels. If a fan or reporter were to use flash-based photography or illuminated video taping, these markers would retroreflect this energy back into their camera causing noticeable bright spots on top of the greyish circles.
To overcome these drawbacks, the present inventors teach the use of ink-based markings that are applied directly to the substrates worn or held by the players. Unlike the spherical markers, these inks will not add any appreciable thickness to the player's clothing or equipment. This effectively overcomes the marker size problem. To overcome the broad band reflectivity drawback, the preferred embodiment includes either reflective, fluorescent or retroreflective inks and compounds that have been engineered to only reflect the narrow band of non-visible tracking energy and as such are substantially transparent to visible light.
C.2. What are the existing ambient electromagnetic energies?
Within the live setting of at least a hockey arena, there will be large area high intensity discharge (HID) lighting such as Metal Halide or Mercury Vapor lamps. These sources typically are chosen because they generate a very broad range of energy throughout the visible spectrum producing a natural-looking white light. The placement of these lamps will be determined and conducive to audience viewing considerations rather than object tracking requirements. During the contest, these lights may be altered in intensity or augmented with additional lighting for visual effect. There is no assurance that the illumination levels in the arena will remain consistent in intensity. Furthermore, it must be anticipated that either fans or broadcasters will bring in additional portable light sources for filing purposes that will also add to the uncontrolled lighting levels.
For beacon-based systems such as Trakus and Orad that depend upon microwave transmissions, the visible light frequencies do not present any noise to the tracking system. Fans using portable cell phones will be transmitting microwave energy but the beacon signals can simply be set to emit at a different wavelength.
For vision-based systems such as Motion Analysis and Vicon, these ambient frequencies are an important source of noise especially in consideration of the various colored uniforms and equipment worn by players during a typical live contest. Hence, if the players are wearing red uniforms or they have red streaks on their jerseys or equipment, then the ambient visible light sources will tend to reflect this color into the tracking cameras. These reflections are similar to those caused by the red LED ring lights as their emitted energy strikes the ice surface. As previously mentioned, the reflections from the ring lights are “mapped out” via software during system calibration. This software technique could not be used to “map out” the dynamically changing red light reflections caused by the player's jersey and equipment colors. It should be noted that these systems include red light filters on their cameras effectively restricting noise to a narrow band overlapping the emitted tracking energies, but not fully eliminating the problem.
As previously mentioned, machine vision companies such as Motion Analysis and Vicon do offer their systems using IR light. This non-visible frequency is used when the customer requires that the tracking be done in darkness. The IR tracking energy is not being used to reduce system noise by cutting down on the visible spectrum.
In order to avoid noise created by the existing visible light sources, the present inventors prefer to work in a non-visible tracking energy such as UV or IR and to employ energy-absorptive techniques elsewhere described in the present and co-pending applications.
C.3. Are their any other pre-existing energy sources that may have available by-product energy that could be used for tracking?
Several of the HID lamps used for large area illumination also generate non-visible frequencies such as UV and IR. Typically, the manufacturers of lights encase the inner bulb and filament with an outer bulb that has been filled with a special vapor mixture and coated with a special compound. The combination of the additional glass bulb, the special vapor and the special coating act to absorb the non-visible energies rather than release them into the surrounding environment. The present inventors anticipate modifying these various light sources to specifically emit some additional portion of their generated non-visible energy that is currently being absorbed.
C.4. What is the natural reflectivity of the various background surfaces to the different potential tracking energies?
Within a hockey arena, the hard mirror-like ice surface presents a challenge to both beacon- and machine vision-based tracking. For instance, other playing surfaces such as grass would tend to absorb most of the ambient frequencies including microwave, red light, UV and IR while only reflecting green light. Another reason the present inventors prefer to track the players and equipment using a non-visible energy such as UV or IR is that the background can then be first treated with either non-visible absorbers, reflectors, or both. This helps to create a clear reflective distinction between the markers and the background thereby facilitating image analysis. The present inventors also prefer to place these non-visible absorbers or reflectors onto the foreground objects such as the players and equipment prior to marking them. In this way, the detectable intensity levels of any unwanted reflections off the foreground objects can be controllably differentiated from the marker's reflections.
C.5. What is the size of the tracking area relative to the range of the potential tracking methods?
Some sports have relatively small playing areas and few players, such as tennis, while other sports have very large areas and many players, such as golf. Sports such as ice hockey, football, soccer and baseball have mid-to-large sized areas with many players to track simultaneously. Microwaves such as those used in the beacon-based systems from Trakus and Orad work well in any of these various-sized playing surfaces. Vision-based systems as currently implemented by companies such as Motion Analysis and Vicon are challenged to track areas significantly larger than a tennis court in size.
The major reason for their difficulties is the “volume tracking” approach they have taken to solving this problem. Simply put, each contestant may move about the playing surface which because of his or her own height and the expected flight of any game objects, becomes a playing volume rather than an area. Every portion of this volume must be in view of two or more cameras at all times. The cameras must be fixed prior to the contest so that they may be calibrated as a network. If players will have a tendency to bunch up, additional cameras will be needed within any given volume to create additional views thereby reducing anticipated inclusions. The cameras must be limited in the field of view so that they can maintain a sufficient resolution or pixels per inch within their field in order to detect the reflected markers. As the playing area widens, it becomes increasingly difficult to place cameras close enough to the inner volumes so that the ideal field of view is maintained per camera without causing an obstruction to the players or viewing audience. This obstruction would occur if a mounting structure were created to hang the cameras directly above the inner volumes. Newer cameras will continue to provide higher resolutions theoretically allowing the cameras to move further back from any given volume and still maintain the requisite pixels per inch resolution. As cameras pull back, however, the distance between the energy tracking source, the reflective marker and the cameras will continue to increase thereby having a negative effect on signal strength.
The present inventors prefer to separate player (object) tracking into two distinct sub-processes thereby eliminating the aforementioned problems. In distinct contrast to the “volume tracking” approach, the preferred embodiment of the present invention relies upon a “player following” controlled by an “area tracking” technique. In essence, the players are first tracked in two dimensions, X and Y, throughout the playing area. The currently determined location of each player is then used to automatically direct that player's individual set of cameras that “follow” him or her about the playing surface. This two-step approach has many critical advantages when faced with tracking objects throughout a larger volume. First, locating the “top surface” (i.e., helmet) of each player in X-Y space for substantially the entire contest is significantly simpler than trying to detect their entire form from two or more cameras throughout the entire playing volume. Second, by placing the player id on their “top surface” (i.e., their helmet and if need be shoulder pads), the system is able to easily identify each player while it also tracks their X-Y coordinates. Third, by controllably directing one or more automatic pan, tilt and zoom cameras to follow each player (and game object) the ideal field of view and maximum resolution can be dynamically maintained per player.
C.6. Is the tracking environment physically enclosed within a building or outside?
The indoor enclosure is most prevalent in the sport of ice hockey as opposed to other major sports such as football, soccer and baseball. This particular requirement has a negative effect on beacon-based systems such as Trakus and Orad since the hard surfaces of the enclosed arena will cause the microwave tracking energy to go through many reflections before finally dissipating. For vision systems such as Motion Analysis and Vicon, the lower ceilings that are typically found in local youth hockey arenas present a different type of problem. Due to the “volume tracking” approach just discussed, these systems ideally require the placement of their tracking camera assemblies above the playing surface or at least off to the side and very near the “tracking volume.” These assemblies include the camera, lens and filter, the ring lights and power supply as well as a small computer processor for initial data analysis and may typically cost around $15,000. In total distance, each assembly may not be more than twenty to thirty feet off the ice. This makes each assembly prone to damage when a puck is either accidentally or intentionally shot at the camera system. The present inventors prefer an enclosed assembly where the camera, lens and related equipment are protected and yet still able to view the ice surface below through a Plexiglas or similarly transparent covering.
D. Referring next to traditional engineering goals, the preferred and alternate novel embodiments must consider at least the following factors:
D.1. The system should be scalable and therefore comprise uniform assemblies that are combinable into a matrix designed to increase tracking coverage in terms of area, volume or the number of players (objects) while still maintaining uniform performance.
In one respect, beacon-based systems such as Trakus and Orad include a uniform assembly in the form of the player's transmitter that is embedded within their helmet. This design is scalable by the number of players since each additional player is simply given an additional transmitter. In order to be scalable in terms of the subsequent signal processing required to locate each transmitter in X-Y space, however, the system will need to encode select transmitters to select receiving towers. The present inventors do not believe that companies such as Trakus and Orad are currently practicing this technique. This concept requires that a capacity be determined for a single set of receiving towers in terms of the number of transmitters that can be tracked in real time. For example, assume that four receiving towers are initially set up within the arena and that they can only adequately track five transmitters in real time. Once a sixth player is outfitted with a transmitter, the entire system will lose its real time processing capability. It is preferable to equip the initial four towers with a first processor that eliminates any received signal that was not generated by one of the initial five transmitters. In this way the initial five players remain adequately tracked in real time. In order to track an additional five players, an additional four towers and processor should be added to the arena. These towers would then be assigned to specifically process only those transmitters with the unique codes corresponding to the second five players. This approach will make the beacon-based system scalable. Otherwise, the entire system will be “player bound” in that it will have a maximum number of total transmitters that can be tracked irrespective of the playing volume.
Similarly, instead of encoding each signal and digitally filtering all incoming signals for selected codes, each maximum group of transmitters could be assigned a different frequency. The filtering process could then be moved to electronics. Another possibility is to coordinate the timing of the transmitted signals from the various groups of maximum transmitters. In essence, each group of transmitters will be broadcasting their locating signal as specified and non-overlapping intervals coordinated with the receiving towers. The important point is to make the system scalable by defining a minimum configuration of tracking apparatus that can simply be repeated to either increase area, or in the case of the beacon approach, increase players tracked.
For vision-based tracking systems such as Motion Analysis and Vicon, the systems are not uniform in their “camera view per volume.” This lack of uniformity works against system scalability and is due to several coincident design factors. First, the camera assemblies are both relatively expensive in their own right and when taken together represent a significant portion of the system price. An attempt is being made to maximize the use of each individual camera's field of view. Second, the field of view of each individual camera is best thought of as a four-sided pyramid where the apex emanates from the CCD array in the back of the camera. In practice, the first several feet of the field of view are not useable for tracking and even so are more than likely outside of the tracking volume for pragmatic clearance considerations. This can be thought of as taking the top off the pyramid. The pixels per marker inch resolution at the nearest versus farthest points in this pyramid can be substantially different and must necessarily be limited by the minimum acceptable resolution to identify a given size marker.
Third, to be scalable, each tracking volume serviced by two or more assemblies should be either square or rectangular in its cross-sectional shape parallel to the ice surface. In this fashion, these volumes could simply be repeated with slight overlaps for calibration purposes in order to create larger and larger tracking volumes. However, this is difficult to accomplish in practice using overlapping, four-sided pyramids. And finally, an individual player with multiple markers attached could transverse any portion of the tracking volume at any rotational angle and it is necessary that each marker remain in the site of two cameras at all times. This requirement calls for a minimum of four cameras surrounding any given volume. In practice six cameras is a more acceptable minimum. As additional players are added to the tracking volume, it is not difficult to see how easy it will be for one player to block the camera's view of another. Keeping in mind the player identification technique practiced by companies such as Motion Analysis and Vicon where the unique “full body” constellation of markers identifies a skater, inclusions take on greater significance.
The task of designing a uniform configuration of acceptable resolution throughout the entire tracking volume while also accounting for the two camera minimum per marker and multiple player bunching is formidable. Furthermore, the end result “wastes” camera field of view since the actual cross-sectional shape formed by the configuration will not be either a square or rectangle nor will it even be uniform in size as it moves from the ice surface up towards the cameras. Hence, if multiple of these “minimum tracking volumes” were to simply be replicated and slightly overlapped, this would result in a considerable loss of valuable tracking region along the edges of the idealized square or rectangular volume. For all of these reasons, companies such as Motion Analysis and Vicon have not approached the larger “tracking volume” as a super-set of smaller volumes in what could be called a scalable approach. For all practical purposes, each unique tracking volume shape and size in combination with the expected number of objects (or players) requires a “custom” layout of tracking assemblies. As more and more players continue to be added to the volume, more and more cameras will be added based upon best judgements for placement. This entire arrangement could be best described as “player bound” as well as “quasi-volume bound.”
As previously discussed, the present inventors prefer the novel approach of separating the (X, Y) tracking of each player or game object as a whole from the (X, Y, Z) determination of each player's critical locations (e.g., joints and body parts). Furthermore, the novel concept of locating all of the identity markings on the “upper surface” of each player facilitates the top-oriented field of view of the (X, Y) tracking assemblies which naturally limits inclusions due to player bunching. Given this separation of tasks, the present invention becomes highly scalable. Each overhead (X, Y) area tracking assembly covers a fixed and uniform square or rectangular tracking area. Furthermore, the pixel resolution per this area remains substantially constant. To cover more tracking area, simply add additional (X, Y) assemblies. Each player that moves throughout this connected tracking area has their helmets (located at a minimum) and ideally also their shoulders. From this information, the player is identified along with pertinent information including orientation, direction of movement, velocity, and acceleration as well as current relative (X, Y) location. Using the current (X, Y) location as well as the direction of movement, velocity and acceleration, the preferred embodiment controllably directs one or more (X, Y, Z) pan, tilt and zoom cameras to automatically follow the given player. Furthermore, by intelligent inspection of the various projected player paths, the system optionally switches (X, Y, Z) cameras from one player to another to best maximize overall tracking performance. By constantly zooming each (X, Y, Z) camera to maximize player size per field of view, a uniform and ideal pixel resolution per marker square inch is maintained. To cover more players, simply add additional (X, Y, Z) pan, tilt and zoom camera sets per added player.
D.2. The system should be minimally intrusive upon the objects to be tracked and upon the surrounding environment especially if that environment is a live setting.
As was previously discussed, beacon-based systems such as Trackus and Orad are environmentally transparent in terms of their microwave tracking energy but are intrusive in terms of their beacon technology. While these beacons are small and can conceivably be reduced further in size, they will still always be practically limited to locations where they can be embedded for both the safety and comfort of the player as well as the protection of the device itself.
As was also discussed, vision-based systems such as Motion Analysis and Vicon are intrusive in three key areas. First, their tracking energies are in the visible spectrum such as red light that is visible to players and fans alike. Second, their markers have substantial size and are both noticeable and prone to accidents in a full contact sport such as ice hockey. And third, these same markers are covered with retroreflective tape that is broad spectrum reflective including visible light. As such, existing rink illumination and portable lighting such as camera flash bulbs will tend to create unwanted and distracting reflections into the players' and fans' views.
And finally, the preferred embodiment of the present invention overcomes each of these intrusions of the surrounding environment through its novel teachings as follows. First, the preferred tracking energy is in the non-visible frequencies such as UVA or near IR. Second, the large spherical markers have been replaced with an “invisible” ink or paint that can be adhered to the many necessary substrates while adding minimal thickness. And finally, this ink or paint, whether it be reflective, fluorescent or retroreflective in nature, is made “invisible” by its characteristic of only reflecting or emitting in the desired non-visible frequencies.
D.3. The tracking signal-to-noise ratio should be maximized.
The microwave-based beacon systems such as Trakus and Orad will have minimal noise due to ambient electromagnetic energies. However, due in part to the enclosed nature of a hockey arena and more importantly to the hard surfaces within the arena such as the ice surface itself, each signal transmitted from a player's helmet will bounce off these surfaces creating significant signal noise problems.
The vision-based systems such as Motion Analysis and Vicon attempt to limit system noise by using a narrow band of tracking energy such as red light with a corresponding camera filter. Unfortunately, by working within the visible spectrum in their preferred approach, they are susceptible to noise created by existing rink lighting, additional portable lighting and the reflections these sources will cause off player jerseys and equipment. And finally, even if they work in the non-visible IR region, both this IR energy and red light will reflect off the ice surface creating false marker reflections that must be “mapped out” via software techniques creating “dead spots” within the tracking images.
The present invention is the only system to employ a combination of tracking energy absorbent and reflective compounds in order to “control” the tracking frequency noise. By using absorbers, unwanted reflections are significantly reduced or eliminated all together. By using reflective compounds, unwanted reflection can be increased to a designated intensity region that is detectably different from the marker signal intensity. These absorbent and reflective compounds can be applied to all of the background and foreground substrates. By first addressing the noise issues, the foreground objects are preferably set at a minimal or zero reflection state while the background including the ice surface, the boards and the glass is preferably set at a narrow intensity range detectably different from the marker intensity.
D.4. Manufacturing and installation costs should be minimized and the resultant system should be simple for the user to maintain and operate.
To the extent that the present invention has been shown to be scalable, its manufacturing and installation parameters are more easily calculated and maximized thereby leading to reduced costs. Furthermore, because of its uniform scalable design, the present invention will minimize any requirements for operator intervention to resolve inclusions, lost player identity, or similar confusion due in general to signal-to-noise problems.
To the extent that the existing beacon- and machine vision-based systems are not scalable, their manufacturing and installation parameters are not as easily calculated and maximized thereby leading to increased costs. Furthermore, because these systems are not uniformly scalable in combination with the larger tracking volume and number of tracking markers associated with 3D multi-player movement, they are susceptible to errors from inclusion and poor signal-to-noise ratios creating a need for operator involvement.
Accordingly, the objects and advantages of the present invention are to:
Further objects and advantages are to provide:
Still further objects and advantages of the present invention will become apparent from a consideration of the drawings and ensuing description.
Referring first to
Within all major components 1000, there are three characteristics of markers that are categorized as follows. The possible Marker: Emission Method 530 is retroreflective 532, reflective 534 or fluorescent 536. The possible Marker: Physical Form 540 is spherical (attached ball) 542 or flat (embedded/applied ink) 544. And finally, the Marker: Reflective Shape 550 corresponding to the Marker: Physical Form 540 is uniform circular 552 or non-uniform multi-shape 554, respectively.
Within components 1000, there are three characteristics of ID (identification markers) categorized as follows. The possible ID: Location 560 is on the full body 562 or top surface of body 564. The possible ID: Encoding Method 570 corresponding to the ID: Location 560 is a unique constellation 572 or encoded markings 574. And finally, the possible ID: Obtained 580 is during game surface tracking 582 or outside of game surface tracking 584. The last group of components 1000 is the Calibration Method 590 that can be either pre-tracking 592 or simultaneously with tracking 594.
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Referring now to
Also shown are a single set of four movable volume tracking cameras 140-a, 140-b, 140-c and 140-d that are for example currently assigned to track the top surface 110 of a player starting when he first enters the playing surface 102 from the entranceway 12 e. Tracking with cameras 140-a, 140-b, 140-c and 140-d continues as the player transverses surface 102 and ceases when player exits surface 102 and enters team benches 12 f. Once within bench area 12 f, area ID & tracking cameras similar to 124 track the player and also reconfirm the player's identity by viewing helmet sticker 109 tm&id. At any time, the player may subsequently leave bench area 12 f and reenter surface 102 where again his motion is tracked by movable volume cameras 140-a, 140-b, 140-c and 140-d. Eventually, the player will either exit the playing area through entrance and exit 12 e or return again to bench 12 f and be tracked and re-identified by the ID & tracking cameras.
Referring now to
The present inventors anticipate that this same technique, although it would not be ideal, could be used to pre-calibrate the scalable area tracking matrix 504 m. In consideration of area matrix 504 m, the relative orientation of each camera 124 is primarily side-by-side with its neighbors, allowing for a small overlap on the edges of its FOV 120 v. Furthermore, the preferred orientation of FOV 120 v is “top down,” rather than the “perspective” view of cameras 126. Given these arrangements, a preferable pre-tracking calibration technique would be to use a traditional calibration plate incorporating a fixed set of markings held at pre-known distances from each other. This plate would then be held in a fixed position facing up at the junctions between every two cameras 124 overlapping FOV's 120 v. Again, using standard techniques well known in the art, each of the two cameras could then be jointly calibrated by computer 160. Proceeding throughout all camera junctions in the same fashion would complete the calibration of the network to itself The only remaining task would be to calibrate the entire matrix 504 m to the playing surface 102, entrance and exit 12 e and team benches 12 f. This could be accomplished by placing a marking at a fixed pre-known location somewhere within each of the areas of surface 102, entrance and exit 12 e and team benches 12 f. Once captured by computer 160 through one or more cameras within matrix 504 m, these markings at pre-known locations would serve to register the entire matrix.
Now referring to
As depicted in
Furthermore, it is expected that additional volume cameras 140 assigned to track the same player 17 will similarly be simultaneously calibrated with camera 124. It should be noted that player 17 may be straddling a boundary between area tracking cameras 124 and as such two different volume cameras 140 may actually be calibrated for the same player 17 by two different area cameras 124. In practice, this is immaterial since the pre-calibration by system 160 of the entire scalable area tracking matrix 504 m can be thought of as creating one large single area (X, Y) tracking camera. Hence, it can be seen that each of the volume cameras such as 140 in the present figure or 140-a, 140-b, 140-c and 140-d of prior figures that are currently assigned to follow player 17 are simultaneously calibrated frame-by-frame to the overhead matrix 540 m. Furthermore, once calibrated the multiple cameras such as 140-a, 140-b, 140-c and 140-d may be used to stereoscopically locate markings on player 17 that are not in view of the overhead matrix 540 m.
Referring now to
Referring now to
The present inventors have listed many absorbers and treatments that may be used especially to absorb UV frequencies in the prior co-pending application entitled Employing Electromagnetic By-Product Radiation in Object Tracking. Such treatments and absorbers are also well known for the IR frequencies to someone skilled in the art. Further, the present inventors have shown that it may also be similarly beneficial to apply energy reflectors rather than absorbers, especially with respect to the background such as player surface 102. What is important is that the intensity of the reflected signal off the tracking marks be clearly distinguishable from any reflections off the background or foreground (player's body and equipment). To gain this clarity of signal differentiation, it may be desirable to either reduce reflections through absorption or to increase reflection through reflective materials. This concept of absorbers and reflectors for the control of the signal-to-noise ratio was not listed in system 1000 as a separate component since the present inventors see it as a beneficial optimization to every possible combination of components listed in 1000.
Referring now to
Referring specifically to
Referring specifically to
Referring now to
Referring now to
Specifically, precisely measured and spaced track markings 104 m have been placed onto stick 104. The exact size, shape and spacing of these markings 104 m are immaterial to the concept being presently taught. Placing these markings upon a rigid object that is used by each player during the game will provide computer 160 with a way to verify its calibration estimates of both fixed area tracking camera 124 and especially movable volume cameras 140. Another possibility is to place similar markings onto the pipes of the goals on either end of the playing surface.
Referring now to
A careful study of the for prior co-pending applications filed by applicants and identified above along with the previously described components 1000 will suffice to teach those skilled in the art how each component may operate as a functional part of a complete system. Therefore, the remainder of this application will focus on distinguishing at a higher level the various possible novel optimized systems according to the present invention along with discussions as to their tradeoffs.
Referring now to
While system 1002 accomplishes the goals of real-time 3D tracking, it has several drawbacks as follows:
Referring now to
The preferred embodiment 1004 accomplishes the goals of real-time 3D tracking without the limitations of currently available system 1002 providing the following advantages:
Referring now to
Embodiment 1006 accomplishes the goals of real-time 3D tracking, without the limitations of system 1002, and provides the following additional advantages:
Referring now to
Embodiment 1008 accomplishes the goals of real-time 3D tracking, without the limitations of system 1002, and provides the following additional advantage: using retroreflective markers that reflect only the narrow band of tracking frequencies of IR or UV, the system provides for greater reflected signal strength while still remaining visibly transparent to both players and audience.
Referring now to
Embodiment 1010 accomplishes the goals of real-time 3D tracking, without the limitations of system 1002, and providing the following additional advantage: identifying players simultaneously with game surface tracking the system provides the option of eliminating separate area tracking cameras in the non-playing surfaces of the entrance and exit passageway and team benches.
Referring now to
Variation 1012 incrementally improves upon the real-time 3D tracking implemented by system 1002 by providing the following additional advantages:
Referring now to
Variation 1014 incrementally improves upon the real-time 3D tracking implemented by non-visible variation 1012 by providing the following additional advantages:
Referring now to
Variation 1016 incrementally improves upon the real-time 3D tracking implemented by top surface of body encoded id variation 1014 by providing the following additional advantages:
Referring now to
Variation 1018 incrementally improves upon the real-time 3D tracking implemented by outside of game surface id variation 1016 by providing the following additional advantages:
Referring now to
Variation 1020 incrementally improves upon the real-time 3D tracking implemented by existing light source variation 1018 by providing the following additional advantages:
Referring now to
Variation 1022 incrementally improves upon the real-time 3D tracking implemented by outside of game surface id variation 1016 by providing the following additional advantages:
At least the following components are considered to be novel and first taught by the present inventors for use within a multi-object tracking system:
There are other novel components first taught by the present inventors that are not specifically identified in any of the
The use of these components is considered to be important and equally applicable to any such system 1002 through 1022.
Thus the reader will see that the present invention successfully:
Furthermore, the reader will also see that, for at least the preferred embodiment and to a great extant its alternates as well as the variations of the systems, the present inventors have taught how to construct a system that:
While the above description contains many details, these should not be construed as limitations on the scope of the invention, but rather as an exemplification of preferred embodiments thereof. Many aspects of the system's functionality are beneficial by themselves without other aspects being present as will be appreciated by those skilled in the art. Furthermore, all of the novel combinations of components taught have anticipated application beyond that of the tracking of live sporting events. Examples of other applications include but are not limited to the tracking of human actors for the creation of animated film sequences, the tracking of human subjects for medical research, as well as other object tracking functions currently preformed by existing systems.
From the foregoing detailed description of the present invention, it will be apparent that the invention has a number of advantages, some of which have been described above and others that are inherent in the invention. Also, it will be apparent that modifications can be made to the present invention without departing from the teachings of the invention. Accordingly, the scope of the invention is only to be limited as necessitated by the accompanying claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US3944738||Dec 11, 1973||Mar 16, 1976||Minnesota Mining And Manufacturing Company||Method to increase the visibility of game objects during telecasting|
|US4189747||Sep 15, 1967||Feb 19, 1980||Hughes Aircraft Company||Infrared tracking system|
|US4539590||Mar 8, 1983||Sep 3, 1985||Gage Richard J||Method and apparatus for processing optical tracking signals|
|US4545576||Aug 13, 1983||Oct 8, 1985||Harris Thomas M||Baseball-strike indicator and trajectory analyzer and method of using same|
|US4664378||Dec 12, 1983||May 12, 1987||Auken John A Van||Electrically conductive tennis ball|
|US4673183||Sep 23, 1985||Jun 16, 1987||Trahan Francis B||Golf playing field with ball detecting radar units|
|US4675816||Aug 26, 1985||Jun 23, 1987||Brandon Ronald E||Electronic method of locating a football|
|US4764668||Nov 25, 1986||Aug 16, 1988||Alcatel Espace||System for locating an object provided with at least one passive target pattern|
|US4866414||Apr 4, 1988||Sep 12, 1989||Sever Diaconu||Optoelectronic lawn tennis linesman system|
|US4905315||Jun 30, 1988||Feb 27, 1990||Solari Peter L||Camera tracking movable transmitter apparatus|
|US4980871||Aug 22, 1989||Dec 25, 1990||Visionary Products, Inc.||Ultrasonic tracking system|
|US5023709||Nov 6, 1989||Jun 11, 1991||Aoi Studio Kabushiki Kaisha||Automatic follow-up lighting system|
|US5082263||Nov 6, 1990||Jan 21, 1992||Richard Berger||Method of and system for determining position of tennis ball relative to tennis court, and tennis ball provided therefor|
|US5091780 *||May 9, 1990||Feb 25, 1992||Carnegie-Mellon University||A trainable security system emthod for the same|
|US5138322||Aug 20, 1991||Aug 11, 1992||Matrix Engineering, Inc.||Method and apparatus for radar measurement of ball in play|
|US5150895||Sep 11, 1991||Sep 29, 1992||Richard Berger||Method of and system for determining a position of ball relative to a playing field, and ball provided therefor|
|US5164827 *||Aug 22, 1991||Nov 17, 1992||Sensormatic Electronics Corporation||Surveillance system with master camera control of slave cameras|
|US5268734||May 31, 1990||Dec 7, 1993||Parkervision, Inc.||Remote tracking system for moving picture cameras and method|
|US5346210||Aug 28, 1992||Sep 13, 1994||Teem Systems, Inc.||Object locator system|
|US5363297||Jun 5, 1992||Nov 8, 1994||Larson Noble G||Automated camera-based tracking system for sports contests|
|US5394824||Mar 8, 1994||Mar 7, 1995||Johnson, Jr.; Lawrence F.||Thermochromic sensor for locating an area of contact|
|US5401026||Jun 29, 1993||Mar 28, 1995||Blackfox Technology Group||Method and apparatus for determining parameters of the motion of an object|
|US5406501||Apr 12, 1994||Apr 11, 1995||U.S. Philips Corporation||Method and device for use in detecting moving targets|
|US5413345||Feb 19, 1993||May 9, 1995||Nauck; George S.||Golf shot tracking and analysis system|
|US5446701||Aug 10, 1993||Aug 29, 1995||Teem Systems, Inc.||Object locator system|
|US5471383||Sep 30, 1994||Nov 28, 1995||Acushnet Company||Monitoring systems to measure and display flight characteristics of moving sports object|
|US5473369||Feb 23, 1994||Dec 5, 1995||Sony Corporation||Object tracking apparatus|
|US5485163||Mar 30, 1994||Jan 16, 1996||Motorola, Inc.||Personal locator system|
|US5489886||Aug 4, 1993||Feb 6, 1996||Alos-Officiating Tennis System Limited||Automatic line officiating system and method thereof|
|US5504477||Nov 15, 1993||Apr 2, 1996||Wybron, Inc.||Tracking system|
|US5508736||Jun 15, 1995||Apr 16, 1996||Cooper; Roger D.||Video signal processing apparatus for producing a composite signal for simultaneous display of data and video information|
|US5564698||Jun 30, 1995||Oct 15, 1996||Fox Sports Productions, Inc.||Electromagnetic transmitting hockey puck|
|US5602760||Feb 2, 1994||Feb 11, 1997||Hughes Electronics||Image-based detection and tracking system and processing method employing clutter measurements and signal-to-clutter ratios|
|US5617335||Jan 18, 1996||Apr 1, 1997||Fujitsu Limited||System for and method of recognizating and tracking target mark|
|US5647015||Jun 7, 1995||Jul 8, 1997||Texas Instruments Incorporated||Method of inferring sensor attitude through multi-feature tracking|
|US5700204||Jun 17, 1996||Dec 23, 1997||Teder; Rein S.||Projectile motion parameter determination device using successive approximation and high measurement angle speed sensor|
|US5731785||May 13, 1994||Mar 24, 1998||Lemelson; Jerome H.||System and method for locating objects including an inhibiting feature|
|US5734343||Jul 18, 1996||Mar 31, 1998||Motorola, Inc.||One-way optical highway communication system|
|US5745126 *||Jun 21, 1996||Apr 28, 1998||The Regents Of The University Of California||Machine synthesis of a virtual video camera/image of a scene from multiple video cameras/images of the scene in accordance with a particular perspective on the scene, an object in the scene, or an event in the scene|
|US5768151||Feb 14, 1995||Jun 16, 1998||Sports Simulation, Inc.||System for determining the trajectory of an object in a sports simulator|
|US5900978||Nov 17, 1994||May 4, 1999||Reflective Technology Industries Limited||Retroreflective materials|
|US5912700 *||Jan 10, 1996||Jun 15, 1999||Fox Sports Productions, Inc.||System for enhancing the television presentation of an object at a sporting event|
|US6061644 *||Dec 5, 1997||May 9, 2000||Northern Digital Incorporated||System for determining the spatial position and orientation of a body|
|US6073489||Mar 3, 1998||Jun 13, 2000||French; Barry J.||Testing and training system for assessing the ability of a player to complete a task|
|US6079862||Jun 18, 1997||Jun 27, 2000||Matsushita Electric Works, Ltd.||Automatic tracking lighting equipment, lighting controller and tracking apparatus|
|US6295367 *||Feb 6, 1998||Sep 25, 2001||Emtera Corporation||System and method for tracking movement of objects in a scene using correspondence graphs|
|US6308565 *||Oct 15, 1998||Oct 30, 2001||Impulse Technology Ltd.||System and method for tracking and assessing movement skills in multidimensional space|
|US6359647 *||Aug 7, 1998||Mar 19, 2002||Philips Electronics North America Corporation||Automated camera handoff system for figure tracking in a multiple camera system|
|US6724421 *||Dec 15, 1995||Apr 20, 2004||Sensormatic Electronics Corporation||Video surveillance system with pilot and slave cameras|
|US6924832 *||Sep 11, 2000||Aug 2, 2005||Be Here Corporation||Method, apparatus & computer program product for tracking objects in a warped video image|
|US20050086704 *||Jun 27, 2003||Apr 21, 2005||Rhodes Scott A.||Video security system|
|EP0300729A2||Jul 19, 1988||Jan 25, 1989||Kabushiki Kaisha Empire Airport Service||Bar code, bar code system and bar code reading method|
|EP0813040A2||Jun 6, 1997||Dec 17, 1997||Xerox Corporation||Precision spatial mapping with combined video and infrared signals|
|EP0959444A1||Aug 7, 1997||Nov 24, 1999||Nurakhmed Nurislamovich Latypov||Method for following and imaging a subject's three-dimensional position and orientation, method for presenting a virtual space to a subject, and systems for implementing said methods|
|FR2726370A1||Title not available|
|GB2191361A||Title not available|
|WO1995010915A1||Oct 11, 1994||Apr 20, 1995||Orad, Inc.||Sports event video|
|WO1997002699A1||Jun 28, 1996||Jan 23, 1997||Fox Sports Productions, Inc.||A system for enhancing the television presentation of an object at a sporting event|
|WO1998046029A1||Mar 18, 1998||Oct 15, 1998||Orad Hi-Tec Systems Limited||Graphical video systems|
|WO1999044698A2||Mar 3, 1999||Sep 10, 1999||Arena, Inc.||System and method for tracking and assessing movement skills in multidimensional space|
|WO2000031560A2||Nov 17, 1999||Jun 2, 2000||Aman James A||Multiple object tracking system|
|1||"Body Speed, Power Stats Going Digital," Street & Smith's Sportsbusiness Journal, vol. 1, Issue 33 (Dec. 7-13, 1998).|
|2||"Gauging The Big Hit," Popular Science, vol. 254, No. 4 (Apr. 1999).|
|3||"How Hard is a Football Hit?", Inventor's Business Daily (Oct. 23, 1998).|
|4||"Much Ado About Nothing," Charles Whipple, pp. 24-25, In Oct. 1998 Photonics Spectra.|
|5||"Stats With Impact," The New York Times, Sec. 6 (Nov. 22, 1998).|
|6||"Trakus Inc.'s Technology Tracks Athletes' Every Move," Boston Business Journal (Oct. 23-29, 1998).|
|7||Ariel Dynamics Inc. product information (Aug. 2001).|
|8||CODA System product information of Charnwood Dynamics Ltd. (Aug. 2001).|
|9||ELITEPlus Motion Analyser product information (Aug. 2001).|
|10||International Search Report Dated Jul. 5, 2000.|
|11||International Search Report Dated Jun. 19, 2001.|
|12||International Search Report Dated Mar. 23, 2000.|
|13||Micromak GmbH product information (Aug. 2001).|
|14||Northern Digital Inc. product information (Aug. 2001).|
|15||Peak Performance Technologies, Inc. product information (Aug. 2001).|
|16||Qualisys North America Home Page-Precision 3D Motion Capture (May 25, 2001).|
|17||RealTime Hires 3D System product information of Motion Analysis Corporation (Aug. 27, 2001).|
|18||Saga-3 RT product information (Aug. 2001).|
|19||Vicon 250 product information of Oxford Metrics Ltd. (at least as early as Feb. 1999).|
|20||Vicon 512 product information of Oxford Metrics Ltd. (Feb. 1999).|
|21||Vicon 8 product information of Vicon Motion Systems Ltd. (Aug. 2001).|
|22||Vicon Cameras product information of Vicon Motion Systems Ltd. (Mar. 2001).|
|23||Visualeyez(TM) Tracking Unit Specifications of PTI 3D Motion-Tracking Systems (Aug. 2001).|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7812842 *||Oct 12, 2010||Sony Corporation||Mobile motion capture cameras|
|US7828478 *||Nov 9, 2010||Delphi Technologies, Inc.||Apparatus and method for thermal detection|
|US7869861 *||Jan 11, 2011||Howmedica Leibinger Inc.||Flexible tracking article and method of using the same|
|US7966636||May 22, 2002||Jun 21, 2011||Kangaroo Media, Inc.||Multi-video receiving method and apparatus|
|US8042140||Oct 18, 2011||Kangaroo Media, Inc.||Buffering content on a handheld electronic device|
|US8051452||Nov 1, 2011||Kangaroo Media, Inc.||System and methods for enhancing the experience of spectators attending a live sporting event, with contextual information distribution capability|
|US8051453||Jul 21, 2006||Nov 1, 2011||Kangaroo Media, Inc.||System and method for presenting content on a wireless mobile computing device using a buffer|
|US8181123||May 15, 2012||Microsoft Corporation||Managing virtual port associations to users in a gesture-based computing environment|
|US8280490 *||Dec 2, 2005||Oct 2, 2012||Siemens Aktiengesellschaft||Registration aid for medical images|
|US8391773 *||Mar 5, 2013||Kangaroo Media, Inc.||System and methods for enhancing the experience of spectators attending a live sporting event, with content filtering function|
|US8391774 *||Jul 21, 2006||Mar 5, 2013||Kangaroo Media, Inc.||System and methods for enhancing the experience of spectators attending a live sporting event, with automated video stream switching functions|
|US8391825 *||Mar 5, 2013||Kangaroo Media, Inc.||System and methods for enhancing the experience of spectators attending a live sporting event, with user authentication capability|
|US8432489||Jul 21, 2006||Apr 30, 2013||Kangaroo Media, Inc.||System and methods for enhancing the experience of spectators attending a live sporting event, with bookmark setting capability|
|US8457719||Dec 8, 2010||Jun 4, 2013||Stryker Corporation||Flexible tracking article and method of using the same|
|US8512160 *||Sep 8, 2003||Aug 20, 2013||Acushnet Company||Multishutter club-ball analyzer|
|US8515669||Jun 25, 2010||Aug 20, 2013||Microsoft Corporation||Providing an improved view of a location in a spatial environment|
|US8547431 *||Jun 8, 2009||Oct 1, 2013||Sony Corporation||Method and apparatus for generating an event log|
|US8666115||Oct 13, 2010||Mar 4, 2014||Pointgrab Ltd.||Computer vision gesture based control of a device|
|US8693732||Jul 10, 2013||Apr 8, 2014||Pointgrab Ltd.||Computer vision gesture based control of a device|
|US8701147||Sep 19, 2011||Apr 15, 2014||Kangaroo Media Inc.||Buffering content on a handheld electronic device|
|US8762894||Feb 10, 2012||Jun 24, 2014||Microsoft Corporation||Managing virtual ports|
|US8842164 *||Aug 26, 2009||Sep 23, 2014||Sony Corporation||Image pickup apparatus and video recording and reproduction system|
|US8854594 *||Aug 31, 2010||Oct 7, 2014||Cast Group Of Companies Inc.||System and method for tracking|
|US8938124||Jun 25, 2013||Jan 20, 2015||Pointgrab Ltd.||Computer vision based tracking of a hand|
|US9015638||May 1, 2009||Apr 21, 2015||Microsoft Technology Licensing, Llc||Binding users to a gesture based system and providing feedback to the users|
|US9046345 *||Apr 30, 2010||Jun 2, 2015||Jan Antonis||Optical probe|
|US9055226||Jan 7, 2013||Jun 9, 2015||Cast Group Of Companies Inc.||System and method for controlling fixtures based on tracking data|
|US9059809||Feb 20, 2012||Jun 16, 2015||Steven M. Koehler||System and method for listening to teams in a race event|
|US9065984||Mar 7, 2013||Jun 23, 2015||Fanvision Entertainment Llc||System and methods for enhancing the experience of spectators attending a live sporting event|
|US9298986||Dec 7, 2012||Mar 29, 2016||Gameonstream Inc.||Systems and methods for video processing|
|US9350776||Jun 15, 2015||May 24, 2016||Tagi Ventures, Llc||System and method for listening to teams in a race event|
|US9350923 *||Jul 21, 2014||May 24, 2016||Cast Group Of Companies Inc.||System and method for tracking|
|US9436998||May 12, 2015||Sep 6, 2016||Leap Motion, Inc.||Systems and methods of constructing three-dimensional (3D) model of an object using image cross-sections|
|US20040147839 *||Oct 25, 2002||Jul 29, 2004||Moctezuma De La Barrera Jose Luis||Flexible tracking article and method of using the same|
|US20050054456 *||Sep 8, 2003||Mar 10, 2005||William Gobush||Multishutter club-ball analyzer|
|US20060067378 *||May 26, 2005||Mar 30, 2006||Rege Siddharth S||Apparatus and method for thermal detection|
|US20060119714 *||Nov 8, 2005||Jun 8, 2006||Sony Corporation||Data processor and data processing method|
|US20060184014 *||Dec 2, 2005||Aug 17, 2006||Manfred Pfeiler||Registration aid for medical images|
|US20070019068 *||Jul 21, 2006||Jan 25, 2007||Marc Arseneau||System and Methods for Enhancing the Experience of Spectators Attending a Live Sporting Event, with User Authentication Capability|
|US20070021056 *||Jul 21, 2006||Jan 25, 2007||Marc Arseneau||System and Methods for Enhancing the Experience of Spectators Attending a Live Sporting Event, with Content Filtering Function|
|US20070021057 *||Jul 21, 2006||Jan 25, 2007||Marc Arseneau||System and Methods for Enhancing the Experience of Spectators Attending a Live Sporting Event, with an Audio Stream Selector Using a Priority Profile|
|US20070022447 *||Jul 21, 2006||Jan 25, 2007||Marc Arseneau||System and Methods for Enhancing the Experience of Spectators Attending a Live Sporting Event, with Automated Video Stream Switching Functions|
|US20070076096 *||Oct 4, 2006||Apr 5, 2007||Alexander Eugene J||System and method for calibrating a set of imaging devices and calculating 3D coordinates of detected features in a laboratory coordinate system|
|US20080034598 *||Aug 14, 2006||Feb 14, 2008||Boccardi Joe S||Positioning system for use in a football game|
|US20080211815 *||Jan 11, 2008||Sep 4, 2008||Demian Gordon||Mobile motion capture cameras|
|US20090094188 *||Oct 3, 2007||Apr 9, 2009||Edward Covannon||Facilitating identification of an object recorded in digital content records|
|US20090245580 *||Apr 2, 2009||Oct 1, 2009||Darryl Greig||Modifying parameters of an object detector based on detection information|
|US20100026801 *||Feb 4, 2010||Sony Corporation||Method and apparatus for generating an event log|
|US20100030350 *||Feb 4, 2010||Pvi Virtual Media Services, Llc||System and Method for Analyzing Data From Athletic Events|
|US20100245546 *||Aug 26, 2009||Sep 30, 2010||Yoshihiko Kuroki||Image Pickup Apparatus and Video Recording and Reproduction System|
|US20100277582 *||Apr 30, 2010||Nov 4, 2010||Jan Antonis||Optical probe|
|US20100281436 *||Nov 4, 2010||Microsoft Corporation||Binding users to a gesture based system and providing feedback to the users|
|US20100281437 *||Nov 4, 2010||Microsoft Corporation||Managing virtual ports|
|US20110069179 *||Mar 24, 2011||Microsoft Corporation||Network coordinated event capture and image storage|
|US20110077510 *||Mar 31, 2011||Jose Luis Moctezuma De La Barrera||Flexible Tracking Article And Method Of Using The Same|
|US20110173235 *||Sep 14, 2009||Jul 14, 2011||Aman James A||Session automated recording together with rules based indexing, analysis and expression of content|
|US20110199487 *||Mar 27, 2008||Aug 18, 2011||Abb Research Ltd.||Method for operating remotely controlled cameras in an industrial process|
|US20110285854 *||May 18, 2010||Nov 24, 2011||Disney Enterprises, Inc.||System and method for theatrical followspot control interface|
|US20120050535 *||Aug 31, 2010||Mar 1, 2012||Gilray Densham||System and method for tracking|
|US20130194427 *||Jan 27, 2012||Aug 1, 2013||Robert Hunter||Systems methods for camera control using historical or predicted event data|
|US20140320667 *||Jul 21, 2014||Oct 30, 2014||Cast Group Of Companies Inc.||System and Method for Tracking|
|USRE43601||Nov 4, 2011||Aug 21, 2012||Kangaroo Media, Inc.||System and methods for enhancing the experience of spectators attending a live sporting event, with gaming capability|
|WO2013008236A1 *||Jul 11, 2012||Jan 17, 2013||Pointgrab Ltd.||System and method for computer vision based hand gesture identification|
|WO2013187937A1||Jan 11, 2013||Dec 19, 2013||Alpine Replay, Inc.||Automatic digital curation and tagging of action videos|
|WO2016054729A1 *||May 8, 2015||Apr 14, 2016||Livebarn Inc.||System and method for optical player tracking in sports venues|
|International Classification||H04N5/225, G06T, H04N5/30|
|Cooperative Classification||H04N5/232, G06T2207/30228, G06T2207/30204, G06T2207/10021, G06T2207/10016, G06T7/2093, G06T7/2033, A63B2024/0056, A63B2225/15, A63B2220/806, A63B2220/40, A63B2220/20, A63B2220/13, A63B2024/0028, A63B2024/0025, A63B24/0021, A63B2102/24|
|European Classification||A63B24/00E, H04N5/232, G06T7/20C, G06T7/20U|
|Nov 30, 2010||AS||Assignment|
Owner name: INTHEPLAY, INC., PENNSYLVANIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:AMAN, JAMES A.;HALLER, WILLIAM R.;REEL/FRAME:025432/0512
Effective date: 20040202
Owner name: MAXX HOLDINGS INC., CANADA
Free format text: SECURITY AGREEMENT;ASSIGNOR:IN THE PLAY, INC.;REEL/FRAME:025414/0124
Effective date: 20101118
|Sep 10, 2012||REMI||Maintenance fee reminder mailed|
|Jan 23, 2013||FPAY||Fee payment|
Year of fee payment: 4
|Jan 23, 2013||SULP||Surcharge for late payment|
|Jun 25, 2015||AS||Assignment|
Owner name: MAXX HOLDINGS, INC., CANADA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:ROBERT H. HOLBER, CHAPTER 7 TRUSTEE FOR INTHEPLAY, INC. MAXX HOLDINGS, INC., TS&E, INC. AND JAMES A. AMAN (COLLECTIVELY, THE "PARTIES");REEL/FRAME:035905/0530
Effective date: 20150618